Article Source

• Categorical Reparameterization with Gumbel-Softmax

Categorical Reparameterization with Gumbel-Softmax

• Eric Jang, Shixiang Gu and Ben Poole
• Chris J. Maddison, Andriy Mnih and Yee Whye Teh

Bayesian Deep Learning Workshop

• NIPS 2016
• December 10, 2016 — Centre Convencions Internacional Barcelona, Barcelona, Spain

Abstract

Categorical variables are a natural choice for representing discrete structure in the world. However, stochastic neural networks rarely use categorical latent variables due to the inability to backpropagate through samples. In this work, we present an efficient gradient estimator that replaces the non-differentiable sample from a categorical distribution with a differentiable sample from a novel Gumbel-Softmax distribution. This distribution has the essential property that it can be smoothly annealed into a categorical distribution. We show that our Gumbel-Softmax estimator outperforms state-of-the-art gradient estimators on structured output prediction and unsupervised generative modeling tasks with categorical latent variables, and enables large speedups on semi-supervised classification.